Abstract
Microsaccades are considered to be involuntary fixational eye movements in addition to a slower random-walk like movement called drift. Following a perceptual event, as well as prior to anticipated events, the microsaccades are typically inhibited and later released in a time course that depends on stimulus properties and cognitive factors. We examined whether the drift also follows the same inhibitory pattern. Method: During fixation observers (n=16) viewed passively a sequence of randomly ordered Gabor patches with varied contrast flashed every 1 second. In a second experiment, observers (n=22) performed a continuous performance task (CPT) with stimuli presented every 2 seconds. The time courses of microsaccade rate, pupil size and drift amplitude were computed relative to stimulus onset based on high speed eye tracking data. Drift was computed after filtering out all traces of saccades, using "box-counting" and the position range within a sliding window. Overall, ~20,000 trials were examined. Results: The drift time course was similar to that of microsaccade rate, though with larger variability. It showed a similar dependence on contrast, and on the history of preceding events. Drift inhibition was correlated with microsaccade inhibition across individuals in the pre-stimulus period of the CPT. Moreover, it persisted in trials with no microsaccades, ruling out an indirect effect of microsaccades on drift. To minimize the possible crosstalk artifact of large pupil size changes on position, we analyzed iso-luminance stimuli as well as pre-stimulus periods. In these conditions, we found no correlation between the pupil derivative and drift, suggesting a genuine drift effect. Conclusion: The drift, like microsaccades and perhaps other types of body movement, appears to show a "freeze effect" in response to perceptual events as well as prior to anticipated events. The pattern of this effect conveys information on the time course of the cognitive processes involved.
Meeting abstract presented at VSS 2014